• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 2649 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS54341: Or the LM14030 for a GSM/LTE Module?

Pradeep Mohan
Intellectual 725 points
Part Number: TPS54341
Other Parts Discussed in Thread: TPS564201, LMR14030, , TPS54260, LM76003, LM73605, TPS54340, LMR23630, LM43603

Hello TI,

I used a TPS564201 (DCAP) DC/DC Convertor with my previous design to power a GSM/LTE Module. The transient response was fantastic. However since DCAP does not offer Wide Vin like current control mode and since the module will be powered by an automotive 12 V battery, I have to reconsider my DC/DC convertor choice. 

1. Are there differences between the transient responses of the LMR14030 and the TPS54341?

2. Why does the Texas Instruments App note SLVA412 use a TPS fmaily device (TPS54260) instead of the LMR family with integrated compensation components and much lower Iq?

3. The device requirements are as follows

4.5 V<VIN<42 V

Iout>3 A

Light load efficiency (Eco Mode)

Smallest footprint possible. 

Which one would be your recommendation?

4. Do you see any advantage using the TPS54341 over the LMR14030? The only I can see if that the lead time is much lower than that of the LMR14030. 

Any suggestions would be much appreciated.

If you feel any other convertor would serve the purpose better, please do point it out. 

Thank You.

  • 0
    TI__Genius 11085 points
    The LMR14030 is internally compensated and the TPS54341 is externally compensated. If both carefully designed, the transient responses should be comparible. The App note was published before the LMR family is released, that's why it was using the TPS family.

    Note that the LMR family max operating Vin is 40V, not 42V. Other than that, LMR should be a good choice for this application.

    You can also consider newly released LM76003, 60V 3.5A synchronous buck. Since it is synchronous, no external diode needed, the total solution size could be smaller than the non-sync solutions. It is also internal compensated with very low IQ.

    Regards,

    Yang
  • 0
    TI__Mastermind 27045 points

    Hi,

    LMR14030 is a better choice as it's newer design and more cost effective than TPS54341.

    Do you need Q-grade for automotive application? 

  • 0 in reply to Andy Chen
    Intellectual 725 points

    Thank you Yang, Thank you Andy!

    The LMR14030 indeed seems to be a better choice. However the lead time seems to be nearly 24 weeks as compared to the 15 of the TPS54341. That is the only reason why I'm still leaning toward the TPS..

    We do not need Q certification.

    Any other recommendations with acceptable lead times?

  • 0 in reply to YangZhang
    Intellectual 725 points
    Hello Yang,

    thank you for the detailed response. The LM76003 seems to be an excellent option too. However the product seems to be really new and I'm worried about the availability, especially for prototypes.
  • 0 in reply to Pradeep Mohan
    TI__Genius 11085 points

    The LM76003 EVM is in stock to order now, www.ti.com/.../lm76003evm-500k

    Sorry we cannot be much helpful when the part is not in stock.

    Regards,

    Yang

  • 0 in reply to YangZhang
    Intellectual 725 points
    Hello Yang,

    thanks for the reply.

    The LM76003 is out of question since we need the chips on our prototypes to be ordered this month.

    I'm also considering the fixed frequency LMR23630A. These seem to be easily available. I've always used >500 kHz switching for our GSM/LTE modules. What would be the disadvantage of 400 kHz? It will still be able to handle the transient loads right?
  • 0 in reply to Pradeep Mohan
    TI__Genius 11085 points
    It will be able to handle transient loads. The drawback of lower frequency is that you will need larger inductance and capacitance for the output. They may lead to larger size components. The benefit of lower frequency is higher efficiency, so you do not need as much space to dissipate the heat. The solution size could be similar.

    If 36V Vinmax parts are ok for your application, the other option is LM73605. It provides adj frequency and pretty good efficiency.

    Best regards,

    Yang
  • 0 in reply to YangZhang
    Intellectual 725 points

    Hello Yang,

    that's a really good point. Thank you.

    I was reading the application note from TI regarding efficiency of synchronous vs non synchronous convertors here. 

    For a typical Vin of 12V, Vout 3.8V, maximum load current 3A, transient loads from I=0.3A to I=3A, and the requirement of minimum current consumption at Iout=3mA when the GSM module is in sleep mode, which would be the best choice?

    How would I evaluate the light load efficiency of one convertor against the other?

    TI's portfolio is really expansive, which is great, but though I've tried my best to filter through the options I think without assistance from this forum I'm going to have a hard time finding the perfect match for my requirements.

    The LM73605 seems more expensive than the LMR26360.

    The final requirement is that I be able to shutdown the switching convertor by pulling the enable pin low and turn it on by pulling the pin high (with V=2.65V). The TPS54340 with its UVLO might not be the best option if I needed to do this (please correct me if I'm wrong).

    This is one reason I'm considering the LMR26360 over the TPS54340. 

    What would you recommend?

  • 0 in reply to Pradeep Mohan
    TI__Genius 11085 points
    Hi Mohan,

    The efficiency is affected by many factors, sync or non-sync, parameters of the internal FET(s) and external diodo if non-sync, the control strategy at light load, how fast the FET(s) turns on/off, etc. To compare light load efficiency among different devices, closed loop no load Iq is one spec to look at, and the efficiency curves on the datasheet can be used.

    Yes the LM73605 is more expensive than the LMR23630, because it is a larger device for 5A max DC load current and it provides more features for design flexibility. Since it is designed to handle larger current, the on-time resistance (RDS-ON) of the internal FETs are lower than the LMR23630, so better efficiency at 3A. Which one to use depends on the design spec and cost target.

    In terms of the turning on and off the device using EN pin, all three choices should work.

    What is the Vout of your design? If it is 5V, I wouldn't suggest TPS54340, since your min Vin is 4.5V. The LM73605 and LMR23639 behave better when Vin very close or even lower than Vout.

    The other adj frequency 36V 3A device is LM43603. It should be less expensive than the LM73605, but more than LMR23630.

    Regards,

    Yang
  • 0 in reply to YangZhang
    Intellectual 725 points

    Hello Yang,

    wow, thanks for the comprehensive reply. That's more information than I could've ever amassed on my looking reading through application notes. Thanks a ton!

    I'll compare the closed loop no load Iq and the efficiency curves. 

    1. The EN pin... We want to be able to control the enable pin of the convertor with another microcontroller output (3.3V). Correct me if I'm wrong, but I cannot do this with the TPS family right? The voltage supplied will set the UVLO threshold if I'm not wrong? Without using a P channel MOSFET I'm not sure I can get it to work that the DC DC convertor is on when the microcontroller output is high and OFF when the output pin is low. The precision enable on the LMR23630 seems more suitable for our application. 

     

    2. Vout of our design is 3.8V, vin should typically be between 9 to 16V. A TVS diode with a reverse working voltage of 24V and a shottky diode will be added before Vin to protect against the test impulses in ISO 7637-2. 

    Please find below the link to our WeBENCH design

    https://webench.ti.com/appinfo/webench/scripts/SDP.cgi?ID=12A3D1341BD602C4

    3. Yes, thank you for the LM43603 suggestion. That is another suitable device, However, this device also has a long lead time. The LM73605 and LM76003 are also devices I did not choose because of the lead time and availability since they seem to be really new devices.

    4. Is any reason you would suggest us not to go with the LM23630? I think this might be the device we will finally select as it meets most of our criteria. There is one thing that worries me though. The steady state simulation shows a curve I'm not familiar with. The ripple is of course completely acceptable, but this pattern with the peaks on the waveform is something I've never seen all the other devices we have worked with. 

    Once again, thanks a ton for your prompt support!

  • 0 in reply to Pradeep Mohan
    TI__Genius 11085 points
    I think the LMR23630 is a good choice for your application. The waveform shown indicates that the output capacitor has rather large ESR. I cannot open the design you've shared. Please check the type of the capacitor used at the output. If ceramic capacitors are used, the waveform shouldn't look at this.
    The step change on the Vout waveform is due to the current going to the capacitor * ESR.